Parent child not working consistently

hi All,

We have been using parent child structure since v5 and migrated to v7.2 earlier in Jan and now upgrading to 7.8.1. After moving to v7.2, we have been seeing inconsistency when using parent-child relationship.

We reload the data on a nightly basis. Once in a while, even though the data load has successfully completed and we are able to see join_field and routing properly set, both has_child and has_parent queries stops working all of a sudden. We are able to query the records independently, when this issue occurs, but not with has_child or has_parent.

If we reload the data again, issue gets resolved. We have been unable to identify what triggers this issue. We have seen this issue happen on both 7.2.0 and 7.8.1.

So I am wondering, if there is anyway to internally verify if the parent child relationship has been actually established?

Thanks!

Yes, we too have a massive data in production of size ~10TB with best compression & looks like hitting the same bug, intermittently the join queries are either not working/taking too long to respond which is not the case before we upgrade the cluster from 7.5.0 to 7.8.1/7.9.2

Since the downgrade is also challenging & not straightforward, not sure how to proceed.

any help from elastic team could be appreciated ASAP....